The heroine with a thousand faces: Universal trends in the characterization of female folk tale protagonists

Jonathan Gottschall

Description

From the book : Evolutionary Psychology 3: 85-103.
Literary scholars and psychologists have long remarked striking similarities in the depiction of male heroes in the world's folk tale traditions.
The best-known attempt to document and explain these similarities is Joseph Campbell's The Hero with a Thousand Faces (1936).
Campbell's work differs in detail from other prominent attempts to define universal features of heroes, however its grand thesis is much the same: while the details of heroes “faces” may change as the investigator crosses geographical, ethnic, cultural and chronological borders, certain details of the hero's life and challenges are everywhere the same.
However, in contrast to the many determined efforts to generalize about heroes, scholars have expended little effort generalizing about cross-cultural features of heroines.
The present article, based on a quantitative content analysis of folk tales from 48 culture areas around the world, represents the first systematic attempt to identify and explain cross-cultural trends in the characterization of heroines.
This study was designed to improve substantially on previous attempts to define literary universals by drawing upon theory and methods used by human scientists to fruitfully explore and explain human universals.

Sujets

• Content analysis
• Evolution
• Evolutionary psychology
• Gender differences

Informations

• Publié par: evolutionary-psychology
• Publié le: 01 janvier 2005
• Nombre de lectures: 46
• Licence : En savoir +
  Paternité, pas d'utilisation commerciale, partage des conditions initiales à l'identique
• Langue: English

Extrait

Evolutionary Psychology  human-nature.com/ep  2005. 3: 85-103 ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯   Original Article   The Heroine with a Thousand Faces: Universal Trends in the Characterization of Female Folk Tale Protagonists  Jonathan Gottschall, First Year Program, c/o English Department, St Lawrence University, Canton, NY 13617, USA. Email: [e-mail].  with fifteen St. Lawrence University undergraduate students: Rachel Berkey, Mitch Cawson, Carly Drown, Matthew Fleischner, Melissa Glotzbecker, Kimberly Kernan, Tyler Magnan, Kate Muse, Celeste Ogburn, Stephen Patterson, Christopher Skeels, Stephanie St. Joseph, Shawna Weeks, Alison Welsh, Erin Welch.  Abstract:Literary scholars and psychologists have long remarked striking similarities in the depiction of male heroes in the world's folk tale traditions. The best-known attempt to document and explain these similarities is Joseph Campbell's The Hero with a Thousand Faces(1936). Campbell's work differs in detail from  other prominent attempts to define universal features of heroes, however its grand thesis is much the same: while the details of heroes faces may change as the investigator crosses geographical, ethnic, cultural and chronological borders, certain details of the hero's life and challenges are everywhere the same. However, in contrast to the many determined efforts to generalize about heroes, scholars have expended little effort generalizing about cross-cultural features of heroines. The present article, based on a quantitative content analysis of folk tales from 48 culture areas around the world, represents the first systematic attempt to identify and explain cross-cultural trends in the characterization of heroines. This study was designed to improve substantially on previous attempts to define literary universals by drawing upon theory and methods used by human scientists to fruitfully explore and explain human universals.  Keywords: Literary Universals, Folk Tales, Content Analysis, Evolutionary Theory, Evolutionary Psychology, Gender Differences, Female Literary Characters.  ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯  Introduction  Literary scholars and psychologists have long remarked striking similarities in the depiction of male heroes in the world's folk tale traditions. The most prominent attempt to document and explain these similarities is Joseph Campbell'sThe Hero

The Heroine with a Thousand Faces

with a Thousand Faces(1936). Campbell's work differs in detail from other notable attempts to define universal features of heroes (e.g., Dundes, 1980; Fontenrose, 1959;  Rank, 1909; Raglan, 1936; Tylor, 1871; Von Hahn, 1876), however its grand thesis is much the same: wherever you travel in the world's folk literatures, heroes will share certain predictable patterns of characteristics; while the details of heroes faces may change as the investigator crosses geographical, ethnic, cultural and chronological borders, certain details of the hero's life and challenges are everywhere the same. However, in contrast to the many determined efforts to generalize about heroes, scholars have expended little effort generalizing about cross-cultural features of heroines. A review of almost 2,000 publications on women and folklore featured not a single concerted attempt to define cross-cultural features of heroines (De Caro, 1983), and we are aware of none that have been published since. The present article, based on a quantitative content analysis of folk tales from 48 culture areas around the world, represents the first systematic attempt to identify and explain cross-cultural trends in the characterization of heroines.  A Word on Terminology  First, the lack of preexisting cross-cultural studies of common traits of heroines is not due to simple, sexist disregard. In fact, a great deal has been written about female folk tale characters (for reviews see De Caro, 1983; DeGraff, 1987; Haase, 2000). The paucity of attempts to generalize about cross-cultural aspects of heroines stems, in large part, from the difficulty of identifying and defining a type of heroine distinctly enough for cross-cultural analysis (see Ragan, 1998 on the difficulties of devising a suitable definition of the heroine). Heroines are not merely heroes with some different body parts; as we shall see later in the paper, there is a distinct scarcity of female characters that are heroic in the same sense we usually associate with male heroes. Therefore, in order to avoid sticky problems of definition, and in order to attain a large and diverse sample, this study defines the term heroine in its most generic sense, as any main female protagonist. As we employ the term, there is absolutely no connotation of special grandeur, virtue, nobility, or courage. The term literary universal must also be defined. The first property of a literary universal is that the putative feature need not appear ineveryliterary work nor must it appear inall An absolute universal is one that appliesthe world's literary traditions. across all literary works and traditions. This is a special, and perhaps rare phenomenon, which likely applies mainly at the highest levels of abstraction (e.g., "all peoples have literature," "all literature is focused on conflict," "all peoples have creation myths," etc.). This article seeks to document features of the characterization of female protagonists that are statistical, not absolute, universals. These are features of characterization which recur at rates significantly greater than would be predicted by chance, and which cannot be convincingly attributed solely to diffusion (see Hogan, 1997, 2004). 

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 86 -

The Heroine with a Thousand Faces

Problems with Previous Attempts to Define Literary Universals  While there is, at present, a resurgence of interest in the subject of literary universals (Arleo, 1997; Carroll, 2001; Gottschall, 2004; Gottschall, Berkey, Drown et al., 2004; Gottschall, Martin, Rea, and Quish, 2004; Gottschall, Callanan, Casamento et al., 2004; Hogan, 1996, 1997, 2004; Jobling, 2001; Mueller, 1993; Richardson, 2000; Sternberg, 2003; Quasthoff, 1996), the majority of literary scholars have, over the last several decades, viewed the very concept of universals with deep suspicion. This is in stark contrast to a long and prominent universalist tradition that argued for distinct regularities in literatures across geographical locations, historical epochs, and levels of cultural complexity. There are both ideological and practical reasons for this state of affairs. We will address only practical reasons, given that ideological reasons have been ably discussed by other authors (e.g., Appiah 1992; Hogan, 1997, 2004). Literary scholars have found it easy to dismiss the concept of literary universals largely because the most prominent research in the field has been rife with theoretical and methodological flaws, and thus vulnerable to devastating skeptical critique (see Cook, 1976; Jobling, 2001). However, the failures in previous attempts to define literary universals may be attributable less to the absence of common aspects of world literatures than to widespread use of weak theory and methods. Prior to the twentieth century universalist claims were based on commonsense claims concerning universal properties of human psychology. For instance, the nineteenth- century anthropologist Edward Tylor wrote:  The treatment of similar myths from different regions, by arranging them in large compared groups, makes it possible to trace in mythology the operation of imaginative processes recurring with the evident regularity of mental law; and thus stories of which a single instance would have been a mere isolated curiosity, take their place among well-marked and consistent structures of the human mind (1871, pp. 281-282).  In the twentieth century, however, psychoanalysis emerged as an explicit and systematic theory of human psychology that attempted to codify and explain the provenance of the universal "mental law(s)" taken for granted by the likes of Tylor. th Psychoanalysis became both the map and the legend for the 20 century's most prominent universalists: the map that indicated where to look for universals; the legend that revealed how to interpret what was found. Thus the most widely known attempts to systematically define literary universals are heir to the shortcomings of psychoanalysis, and have often been guilty of implausible attempts to cram stories into Freudian or Jungian molds. In addition to theoretical weaknesses, previous attempts to define literary universals lacked methodological rigor. Conclusions were not based on statistical analyses of representative samples of texts but on highly

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 87 -

The Heroine with a Thousand Faces

impressionistic "readings" of handpicked texts, typically over-representing circum- Mediterranean content (Cook, 1976; Jobling, 2001).  The research presented in this article represents an attempt to improve upon previous research in literary universals by applying theory and methods that have invigorated the search for, and interpretation of, universals in other fields. By applying scientific methods of data acquisition and analysis an attempt is made to avoid (insofar as is possible) some of the problems of subjectivity, selection, and confirmation bias characteristic of previous attempts. By taking evolutionary theories of human behavior and psychology as map and legend for this exploration, an attempt is made to identify universals that prove more durable and reliable than those based on psychoanalytic theory and looser methodology.  Expectations   This research seeks universal patterns in dimensions of female protagonist characterization where evolutionary theory and research suggests one should find them. The expectations of this study were as follows. On the basis of kin selection theory (Hamilton, 1964) it was expected that female protagonists would devote substantial effort to assisting their kin, especially their close kin, relative to non-kin and distant kin. On the basis of research into human mate preferences inspired by sexual selection theory, it was predicted that female characters, relative to their male counterparts, would place greater emphasis on a potential mate's wealth, status, and kindness (a potential signaller of commitment) than on his physical attractiveness (Buss, 1989). On the flip side, given the heavy emphasis males place on the attractiveness of potential mates in world cultures, it was expected that there would be markedly greater emphasis on the physical attractiveness of female characters relative to male characters. On the basis of Darwin-Trivers sexual selection theory (Darwin, 1871; Trivers, 1972), it was predicted that female protagonists would be identified as less "active," less "courageous," and less likely to be defined as "physically heroic" than their male counterparts. This is because sexual selection theory predicts that, in most sexually reproducing animals, males will be more prone to risk taking behavior in the competition for mates; males' higher likelihood both of reproducing prolifically and dying without issue gives them positive and negative incentives to compete intensely and riskily for mates, and for the social status and resources required to attract and retain them (for reviews of sexual selection literature see Anderson, 1994; Miller, 1999). Finally, it was expected that one side effect of the higher activity, courage, and heroism ratings of males characters would be an abundance of male main characters relative to female main characters. It was assumed that active characters (as well as the courageous and physically heroic) would be more compelling than passive characters, and thus more likely to play central roles in narratives. While some of these expectations may seem obvious on the basis of commonsense, they are at odds with the dominant humanities models, which predict strong inter-cultural variability given the basically arbitrary nature of

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 88 -

The Heroine with a Thousand Faces

human social and gender arrangements.  Data and Methods  Folk tale collections were chosen so as to maximize the geographical variability of the sample as well as variability in levels of cultural complexity. Specialized collections focusing on specific themes, plots, or character types (e.g., Hopi Trickster Tales,Hero Tales of the South Slavs) were rejected in favor of generic collections (Hopi Folk Tales,Traditional Tales of the South Slavs collections). All were of traditional tales, originally transmitted through the oral tradition. In all, the study includes tales from 48 different culture areas from all inhabited continents, varying widely in ecology, geographic location, racial and ethnic composition, political systems, religious beliefs, and levels of cultural complexity. All non-English tales had been translated into English, and the sample ran the gamut from polished fairy tales to literal transcriptions of tales told in traditional contexts. A list of the collections and tales coded can be accessed at this URL or by contacting the lead author: http://www.science.mcmaster.ca/psychology/ehb/gottschall(folktales).pdf.  Once suitable collections were identified, each data collector (10 female and 5 male undergraduates at St. Lawrence University) scanned the thirty longest tales from each of 3 culture areas. (Several researchers were responsible for tales from four culture areas. They volunteered to code collections of tales that only arrived through interlibrary loan after the main portion of the study had been completed.) The coders were participants in a seminar focusing on content analysis methodology and the depiction of female characters in world folk tales. Coders were told that the studys aim was to determine whether or not there were cross-cultural patterns in the depiction of female characters. They were told that either findingeither of prevailing cross-cultural regularity or of prevailing variabilitywould be an important contribution. They were not made aware of the specific theory-derived expectations. The thirty longest tales were chosen, rather than a random selection, to ensure a sample consisting of long, information-rich tales rather than short, information-poor tales. Since the emphasis of the inquiry was on the attributes of female characters, each of 1,440 tales was then scanned for the presence of a main female protagonist or a main female antagonist. A protagonist was defined as a character who plays a central role in the action and who the audience is led to root predominantly for rather than predominantly against. An antagonist was defined as a character who plays a central role in the action, who acts as an obstacle to the goals of the protagonist(s), and who the audience is led to root predominantly against rather than predominantly for. Any tale containing one or both of these character types was flagged for coding. This sampling procedure likely introduced some biases into the sample that will be discussed in the results section.  A coding form was developed to collect data on the main characters of all flagged tales (see appendix 1). While male characters were not the main targets of

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 89 -

The Heroine with a Thousand Faces

the study, coding forms were also filled out for all main male protagonists and antagonists in order to provide a yardstick for measuring the characteristics of the females. Information was gathered on female antagonists for the same reason. In all, researchers completed 1,307 coding forms on 658 different tales. Of the completed forms, 568 were filled out for female protagonists, 392 for male protagonists, 197 for female antagonists, and 150 for male antagonists.  Data analysis is reported for the sample as a whole and for each of 6 geographical regions: 1) the circum-Mediterranean, 2) East Eurasia, 3) North America, 4) South America, 5) Africa, and 6) the Insular Pacific combined with South East Asia and the Pacific Rim (see appendix 2). Where possible our division of culture areas into geographical regions follows Murdock's precedent (1957, 1981), although it diverges in two significant ways. First, the analysis makes up for a shortfall of tales in Murdock's Insular Pacific region and a glut of tales in his East Eurasian region by grouping several culture areas in South East Asia and the Pacific Rim with tales from the Insular Pacific. Second, East African tales were grouped with Africa rather than with circum-Mediterranean tales. These divergences from Murdock result from the necessity of establishing regional samples of roughly equivalent size. The sample was also divided into two broad levels of cultural complexity. The first level consists of tales that circulated primarily in unassimilated band and tribal societies, though the tales may have only been written down after assimilation. The second level consists of tales that, while they may have originated in non-state societies, circulated for long periods in preindustrial state societies. Since the line between these categories can be fine, tales from culture areas that could not be confidently placed in one of these two categories were excluded from the calculation. Finally, because a prominent claim of feminist scholars of folk and fairy tales, is that patterns of characterization in European folk and fairy tales are the result of the patriarchal biases of male editors and collectors (for a survey of these claims see Haase, 2000; for a disconfirming test of these claims see Gottschall, in press), sub- analyses were performed for male-edited collections versus female-edited collections, and for data gathered by female coders versus data gathered by male coders. While there is variation across male and female editors and coders, this variation is not always in a predictable direction and it is consistent with natural variations across collections and cultures. In no case did results for male or female coders and editors violate the general patterns apparent in the other samples (for a report of these results see Gottschall, in press).  Content analysis of literary works presents special challenges to reliability assessment due to the length and complexity of the works and the time consuming nature of the coding (for further discussion see Gottschall, in press). The following conventional measures (see Krippendorff, 1980; Neuendorf, 2002; Weber, 1990) were undertaken to promote inter-coder reliability: we developed a coding question that was as uncomplicated as possible, we devised simple coding instructions, and we prepared a coding dictionary with definitions of all potentially ambiguous terms.

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 90 -

The Heroine with a Thousand Faces

Finally, inter-coder reliability was assessed in two formal tests. The first test assessed the coders' agreement in scanning the same assortment of 23 culturally diverse tales for the presence of a main female protagonist or a main female antagonist. The second test assessed agreement in coding an assortment of 11 culturally diverse tales previously identified as containing either a main female protagonist or antagonist. Most content analysis practitioners strive for reliability rates of 80% or better, and consider 70% to be the minimum level of adequacy (Krippendorff 1980; Neuendorf 2002; Weber 1990). Inter-coder agreement in scanning tales was 89% while, for the variables discussed in this paper, agreement ranged from 75-94% with an average agreement of 88% (see results section for percent agreement on all variables). These rates of reliability were nearly identical to those produced by 12 coders in a different study who applied the same coding question to a western literature sample (see Gottschall, Martin, Quish, and Rea, 2004). This approach, in which reliability ratings are established prior to actual coding rather than having multiple coders read and code all or some fraction of the different works, is well established in content analysis and is discussed as an option in books on, and practical guides to, the subject. The advantage of this approach is that is allows the compilation of the largest data sets, though some content analysts argue that it does so at the cost of some degree of precision in reliability reporting (for discussion of different methods of reliability testing see Krippendorff, 1980; Neuendorf, 2002; Weber, 1990). We acknowledge that the relatively small number of tales examined in our reliability assessment influences confidence in the reliability of our overall results. At the same time, however, we believe that confidence in our results should also be influenced, in a positive way, by the fact that 15 naïve coders, working in isolation from one another, all found the same general patterns of female versus male characterization in their individual collections. While the degree of confidence generated by this fact cannot be expressed quantitatively, such regularity would not be anticipated in the output of unreliable coders.  Results: The Heroine's Face  This study was designed in hopes of producing an empirically sound composite of some broad lineaments that are relatively constant in the heroine's thousand faces. Analysis of data reveals salient trends of female protagonist characterization across the sample in parameters associated with age, levels of physical attractiveness, frequency of representation as the main character, marital status, mating preferences, motivation, level of activity, personality descriptors, propensity for physical heroism, and patterns of altruism.  Frequency of representation as the main character   Across subsamples female protagonists were significantly under-represented as main characters. The representation of female protagonist main characters was

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 91 -

The Heroine with a Thousand Faces

estimated in two ways. The first estimation was reached by subtracting the number of tales with female protagonist main characters in the sample from the total number of tales scanned (all pertinent questions from the coding form are listed in appendix 1). Overall, male main characters outnumbered female main characters by a factor of three (inter-coder reliability 81%). However, this method may overestimate somewhat the representation of male protagonist main characters because it does not account for tales where main characters (like animals, plants, or forces of nature) may be represented as unsexed. In the second estimation, the senior author analyzed the tables of contents of all collections utilized in the study, determining the relative percentages of titles referring to male main characters versus female main characters. Any title that did not communicate definite information as to the sex of the tale's main character was excluded from the calculation. Overall, male main characters outnumbered female main characters by more than 2 to 1. However, this method of estimation likely underestimates the representation of male main characters because it does not account for the fact that most folk tales featuring anthropomorphized animals or forces of nature sex them as male. As indicated in Table 1, the results of both analyses (the first somewhat aggressive, the second somewhat conservative) confirm that female protagonist main characters are significantly under-represented in the samples.  Table 1: Percentage of male and female main characters: two methods of estimation 



  Overall North America South America Circum-Mediterranean East Eurasia Africa Insular Pacific, etc. Bands / Tribes Preindustrial States

 Male 75 (n=1084) 82 (n=247) 84 (n=176) 77 (n=207) 74 (n=222) 62 (n=131) 73 (n=132) 76 (n=567) 79 (n=573)

 Female 25 (n=356) 18 (n=53) 16 (n=34) 23 (n=63) 26 (n=78) 38 (n=79) 27 (n=48) 24 (n=183) 21 (n=147)

 Z-Score 31.45** 20.77** 18.81** 14.65** 13.4** 5.24** 10.01** 23.08** 26.87**

 Male 68 (n=949) 76 (n=151) 76 (n=105) 67 (n=267) 67 (n=216) 68 (n=172) 58 (n=76) 68 (n=441) 68 (n=511)

 Female 32 (n=443) 24 (n=52) 24 (n=34) 33 (n=134) 33 (n=107) 32 (n=82) 42 (n=56) 32 (n=205) 32 (n=235)

 Z-Score 20.59** 11.26** 9.91** 9.96** 9.11** 8.54** 2.49* 14.11** 15.38**

N is number of story characters * indicatesp< .05 indicates **p< .001  (See Appendix for listing of cultures in each region and level of cultural complexity)

 As stark as the discrepancies are, the likelihood exists that, due to the particular sampling methodology we employed, our figures significantlyoverestimate the representation of female main characters. This is because all tales coded were first pre-selected for the presence of prominent female characters. A sample that was not thus presorted would likely have an even more extreme imbalance between main

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 92 -

The Heroine with a Thousand Faces

male and female characters.  These results have recently been replicated in computerized content analyses of a (different) culturally and geographically diverse sample of 43 collections of folktales from around the world. It was discovered that, in these collections, male subject pronouns (he, him, himself, etc.) outnumbered their female equivalents by a ratio of 3:1, and that clearly male terms (man, boy, uncle, etc.) outnumbered their female equivalents by roughly the same margin (Gottschall, Callanan, Casamento et al., 2004.). Together with the results of the present study, these findings suggest that an under-representation of prominent female folk tale characters is statistically, and perhaps absolutely, universal in world folk literatures.  Age  Young, reproductive aged women are vastly over-represented in the ranks of female protagonists (inter-coder reliability 75%). Across subsamples an overwhelming majority of female protagonists were identified either as sexually mature teenagers or as twenty to twenty-nine year olds; just 8% were identified as forty or older. This is in stark contrast to the depiction of female antagonists, 40% of whom were identified as forty or older (see Table 2). Incidentally, male protagonists, though rated slightly older than female protagonists overall, demonstrated exactly the same pattern: the antagonists are significantly older than the protagonists.  Table 2: Percentage of female protagonists and antagonists falling in given age categories   Sexually Mature Teen or Twenties Forty or Older   Female Female Female Female  Protagonist Antagonist Z-Score Protagonist Antagonist Z-Score Overall 80 (n=531) 38 (n=174) 10.3** 8 42 -8.68** North America 71 (n=76) 38 (n=26) 3* 18 35 -1.57 South America 71 (n=75) 52 (n=23) 1.58 8 30 -2.22* Circum-Mediterranean 91 (n=121) 27 (n=44) 8.83** 4 64 -7.96** East Eurasia 85 (n=117) 44 (n=41) 4.83** 4 29 -7.63** Africa 62 (n=78) 24 (n=25) 3.69** 12 48 -3.43** Insular Pacific, etc. 94 (n=63) 53 (n=15) 3.04* 2 33 -2.59* Bands / Tribes 70 (n=263) 41 (n=80) 4.64** 13 36 -3.98** Preindustrial States 88 (n=252) 38 (n=88) 9.12** 3 44 -7.61** 



 

N is number of story characters * indicatesp< .05 ** indicatesp< .001 (See Appendix for listing of cultures in each region and level of cultural complexity)

We did not predict the age discrepancies in the portrayal of antagonists and

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 93 -

The Heroine with a Thousand Faces

protagonists of both sexes and will not speculate here about its root causes. What seems clear, however, is that the youthfulness of protagonists is connected to the prominent role of marriage in the tales.  Marriage and mating preferences  While there was significant variability across sub-samples, the majorities of male and female protagonists were unmarried at the beginning of their tales (77% male; 78% female) and, of these characters, most were married by the end (64% female; 64% male). Fewer previously unmarried antagonists succeeded in marrying by the end of their tales, but the percentage was still substantial (33% female; 22% male). Thus a large proportion of folk tales dwell on the process of attracting and securing mates. As a result, the tales convey information regarding the attributes characters from the world's folk tale traditions value in potential mates. Consistent with expectations, female protagonists were rated as placing the highest premium on a potential mate's kindness (a potential signaller of commitment), followed by his wealth and social status. In contrast, while male characters were also attracted to kindness, they were far more likely to be rated as placing a premium on the physical attractiveness of potential partners, and placed markedly less emphasis on wealth and status (see Table 3) (for further discussion of mate preferences in the tales see Gottschall, Martin, Rea, and Quish, 2004). Inter-coder reliability for the marital status variables were, respectively, 89% and 86%; reliability for the mate preferences variable was 87%.  Table 3: Percentage of male and female characters identified as placing primary emphasis on given mate preference criteria

  Male Female Z-Score Male Female Z-Score Male Female Overall Folk Tales 56 (n=246) 23 (n=278) 7.78** 9 26 -5.03 35 51 N. America 48 (n=34) 6 (n=34) 4.19** 3 39 -3.81** 48 54 S. America 65 (n=48) 51 (n=39) 1.26 13 23 -1.28 22 26 Circum-Mediterranean 42 (n=68) 15 (n=77) 3.68** 6 15 -1.68 52 70 East Eurasia 58 (n=47) 24 (n=59) 3.71** 11 27 -2.25* 30 49 Africa 63 (n=16) 35 (n=20) 1.96* 19 40 -1.45 18 25 Insular Pacific, etc. 84 (n=19) 13 (n=24) 6.67** 5 38 -2.90** 11 50 Bands/Tribes 67 (n=114) 33 (n=94) 5.01** 9 34 -4.47** 25 33 Pre-industrial States 50 (n=115) 17 (n=140) 5.74** 8 21 -3.18** 43 61   N is number of story characters * indicatesp ** indicates< .05p< .001  (See Appendix for listing of cultures in each region and level of cultural complexity)  

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

Z-Score -3.51** -0.61 -0.29 -2.28* -2.08* -0.3 -3.18** -1.89 -3.05** 

- 94 -

The Heroine with a Thousand Faces

Level of physical attractiveness  Consistent with expectations, when information was available regarding a female protagonist's physical appearance, they were almost universally described as physically attractive (results are based on explicit judgments of narrators and characters, not on individual coder preferences). In fact, the appearance of a female protagonist explicitly defined as physically unattractive was a true statistical anomaly: just eight of 1,440 tales included a main female protagonist explicitly defined as unattractive. While male protagonists were also overwhelmingly more likely to be defined as physically attractive than unattractive, information on this factor was much less likely to be conveyed if the character was male. Overall, 50% of tales featuring female protagonists contained explicit information on their physical attractiveness versus just 21% for male protagonists. Further, when there was information available on a characters physical attractiveness, there were approximately 50% more references per tale to female physical attractiveness than male. The female protagonist's physical attractiveness was also emphasized significantly more than that of the female antagonist.  Table 4: Physical attractiveness of male and female protagonists

  Overall N. America S. America Circum- Mediterranean East Eurasia Africa Insular Pacific, etc. Bands/Tribes Preindustrial States

 Male

22 (n=392) 7 (n=72) 13 (n=86) 38 (n=97) 22 (n=65) 24 (n=33) 19 (n=27) 13 (n=192) 29 (n=188)

 Female

51 (n=568) 39 (n=83) 37 (n=86)

64 (n=126) 63 (n=97) 47 (n=83) 45 (n=66) 37 (n=297)

64 (n=267)

 Z-Score

-9.83** -5.16** -3.85**

-4.01** -5.84** -2.46* -2.79* -6.48**

-7.85**

 Male

96 (n=83) 100 (n=5) 85 (n=12)

100 (n=37) 83 (n=17) 100 (n=8) 100 (n=4) 96 (n=25)

98 (n=55)

 Female

97 (n=292) 91 (n=32) 97 (n=32)

95 (n=84) 100 (n=75) 97 (n=39) 100 (n=30) 95 (n=110)

96 (n=169)

 Z-Score

-0.23 1.82 -1.21

2.05* -1.91 1.01 0 0.12

0.98

 Male 1.95 (x=78) 1.3 (x=3) 2.23 (x=13)

1.97 (x=32) 1.86 (x=14) 2.13 (n=8) 1.3 (n=3) 2.09 (n=23)

1.9 (n=52)

 Female 3.03 (x=293) 3.03 (x=30) 2.55 (x=33)

2.93 (x=76) 3.77 (x=60) 3.05 (39) 2.25 (n=32) 2.84 (n=111)

3.08 (n=171)

 T-Stat -5.08** *** ***

-2.56* -3.37** *** *** -1.85

-4.39**

 N is number of story characters * indicatesp indicates **< .05p< .001 ***sample size too small to calculate significance X is number of references to attractiveness per sub-sample (See Appendix for listing of cultures in each region and level of cultural complexity)   Overall, in cases where information was conveyed on physical appearance, just 69% of female antagonists were defined as physically attractive. Moreover, as with male protagonists, information on physical attractiveness was conveyed less frequently (23% of tales) and less repetitively (three times fewer references per tale) for the female antagonist. There was least emphasis on the attractiveness of male

Evolutionary Psychology  ISSN 1474-7049   Volume 3. 2005.

- 95 -